Professional Documents
Culture Documents
net/publication/220961952
CITATIONS READS
28 2,777
2 authors:
Some of the authors of this publication are also working on these related projects:
All content following this page was uploaded by Fabio Petrillo on 14 November 2016.
9
neering found in process of game development, from analysis cinema [17]. The conception of a game can also count on a
of the literature on game development and mainly through higher abstraction level, organized in the Concept or Pro-
the analysis of postmortems. Section 4 discusses the results posal Document. This document can analyze aspects such
of these analysis. Finally, the conclusions are presented in as market, budget, schedule, technology, art style, profile
section 5. of the development group and some high level description
of the gameplay. However, the preparation of the Concept
2. THE TRADITIONAL GAME DEVELOP- Document is not common to all projects [24].
MENT
The development process used in the production of most
The game development community has a vast specialized
games is based on the waterfall model [25]. This process con-
literature, especially about technical issues. The industry
sists of phases that are executed sequentially, in which each
also has a series of titles that describe how to teach and
one generates a product and is independent from the oth-
work in a games company, such as Gershenfeld et al. [12].
ers. The features inherent in producing games require some
adjustments to the classic process, which can be showed in
However, not many software engineering works are dedi-
figure 1.
cated to the gaming industry: two remarkable exceptions are
Bethke [4] and Schofield [11]. It is interesting to emphasize
Some authors argue that the waterfall model, while serving
that these two works are related to a waterfall philosophical
as a common denominator between the cycles exist, not to
viewpoint, in particular the work of Bethke [4], one of the
be fully implemented. Typical features of a game, as
most cited by the games community, demonstrating the ex-
dynamics of design or difficulties in planning the gameplay,
isting culture of prescriptive and non-iterative procedures.
make it difficult - or impossible - to specify completely a
Bethke [4] explicitly advocates the adoption of a subset of
game without writing any code, that is, without a version
the Unified Software Development Process as the develop-
of the system in which the project can be tested. This re-
ment process for electronic games, for the simple fact that
sults in development cycles that involve links between the
this is a “standard” in the software industry.
stages of specification of the game and test procedures. [25]
quotes a cycle of incremental development, the Staged De-
According to [10], it is possible to identify a common cycle
livery Model, as an alternative to the waterfall model.
of development in different teams: basically, the waterfall
model [23] adapted to produce games. In particular, two
stages of this model may be highlighted: a) defining the 3. PMA TO IDENTIFY GOOD PRACTICES
game rules and b) the production of Document Set or In this section, we describe how we have gathered evidences
Project Document. of adoption of good practices in game industry. The Post-
mortems Analysis (PMA) is a technique borrowed from Em-
The rules of a game determine the behavior and interaction pirical Software Engineering. Empirical Software Engineer-
among the characters and their environment, and can be ing aims to investigate and collect relevant data to generalize
viewed as requirements. In fact, game designing is nothing the results, learning from mistakes and successes, providing
more than creating a set of rules [9], usually defined in an a future reuse of this knowledge.
informal process, involving all members of the development
team [26]. Controlled experiments, case studies and surveys approaches
are most commonly used and better known than the PMA
The Project Document (( Design Document)) is the main, [29, 33, 28]. The main reason for the PMA not be adopted in
often the only, documentation of a game. Your goal is to de- software projects in different fields of application is not their
scribe and detail the mechanics of the game, in other words, difficulty of operationalizing, neither the resources involved
what the player is able to do within the environment of the or much less cost, but rather the absence of postmortems.
game, as he is able to do it and how it can lead to a satis-
factory experience. It also usually includes the main com- The term postmortem designates a document which sum-
ponents of the story and describes the scenarios in which marizes the project development experiences, with a strong
the game is set, supporting the description of the player’s emphasis on positive and negative aspects of the develop-
actions. Many developers refer to it as a functional specifi- ment cycle [14]. It is commonly done right after the project
cation, using it as a basis for the architectural design [24]. finishes, by managers or senior project participants [7]. In a
software engineering viewpoint, postmortems are important
The creation of a solid project document is considered, tra- tools for knowledge management [5], from which the group
ditionally, as the most important step in the game develop- can learn from its own experiences and plan future projects.
ment. The difficulties in creating this document are caused, In fact, the postmortem analysis can be so revealing that
mainly, by the nature of the task and the tools used in their some authors [5] argue that any project should be finished
conception, it is not possible, for example, document the without it’s own postmortem.
“fun” [19]. The development team could use his experience
and intuition in defining the mechanics of the game, but the Traditionally, IT teams have no habit of creating postmortems,
quality of entertainment provided by the game in general although there are obviously exceptions (see for example
can only be assessed in stages of testing. [31]) and strong recommendations to do it [5]. Fortunately,
a rare exception is the domain of games.
The game modeling is not changed in recent years and con-
tinues based on narrative techniques, like scripts and story- Postmortems are much used in the game industry. Many
boards, borrowed from other entertainment media, such as game websites devote entire sections to present these doc-
10
uments, such as Gamasutra (http://www.gamasutra.com) length of development, release date, target platforms, and
and Gamedev (http://www.gamedev.net). It is also very the hardware and software used in the development.
interesting to note the variety of development teams pro-
files and projects behind these documents, varying from few The information contained in the postmortems constitute
developers in small projects to dozens of developers in five- knowledge base that can be reused by any development team,
year-long projects. which includes examples and real life development experi-
ences. They can help in knowledge sharing, and can be very
The postmortems published by Gamasutra mainly follow the useful for planning future projects.
structure proposed by the Open Letter Template [20], which
is composed by three sections. The first section summarizes Thus, the 20 postmortems were examined to find good prac-
the project and presents some important aspects of the de- tices. The process of recognizing the good practices took
velopment. The next two sections, however, discuss the most place in 3 phases. At first, each postmortem has been read
interesting aspects to the game developers: and the quotes that were deemed relevant were highlighted.
In the second stage, based on traditional and agile software
engineering knowledge [1, 23, 3], the practices to be tabu-
• What went right: it discusses the best practices lated were selected. In the third stage, each report was read
adopted by developers, solutions, improvements, and again, with special attention to highlighted passages, and
project management decisions that have improved the each citation was tabulated according to the classification
efficiency of the team. All these aspects are critical previously made.
elements to be used in planning future projects.
• What went wrong: it discusses difficulties, pitfalls, To perform the postmortems analysis and the organization
and mistakes experienced by the development team of good practices, we need to select some postmortems , ana-
in the project, in both technical and management as- lyze them and compile the results, forming a set of practices
pects. that are approved or disapproved by these developers. We
performed the same procedure (analysis of literature and
analysis focused on the postmortems) to carry out a diag-
The postmortem is closed with a final message from the au- nosis of the major problems and difficulties encountered in
thor, commonly followed by a project technical brief, which the development of games (see [21]).
includes the number of full- and partial-time developers,
11
Among the various existing postmortems, we selected 20 20) the occurrence of continuous integration found the
postmortems posted on Gamasutra, listed in tables 1 and analyzed projects.
3. The 20 postmortems analyzed were selected randomly,
having one important criterion for the selection: we se- If we calculate the average number of good practices adopted
lected postmortems of projects that resulted in a complete in the analyzed projects, we will see that it is 7.2 prac-
game and delivered to market, not been analyzed reports of tices, with standard deviation of 2.6. This means that most
projects failed, canceled or terminated without a product. projects employed between 5 and 11 practices, and that the
average is about half of the examined practices (7 to 13).
In preparation stage, we studied the main agile practices [8, This result shows that the game industry adopts a consid-
1, 3, 22] and good practices in the gaming industry [27, 13]. erable number of good practices in their projects, undoing
In this study, 12 good practices were listed for analysis. a little pessimistic view of some authors, as [4] and Flood
[10].
The 20 postmortems were read and sentences quoting prac-
tices were highlighted. During this stage, another practice If we analyze these results, we can see that in general, good
was found (“Belief in the success of the project”), which was practices adopted in the traditional software indus-
added to the analyzed set, totaling 13 practices, which are try are also found in the games industry. In both
listed in Tables 1 and 2 and also in Figure 2. Finally, a ta- studies, the quality of the team was dominant for the suc-
ble for data collection was organized (table 1), with projects cess of the project, as well as the concern with the adoption
arranged in rows and practices in columns. of good practices for programming.
Despite these results being not be surprising, they show A new analysis can be offered to collate the results presented
the importance of the team for the project success, con- by [21], which outlines the problems encountered with the
firming the statements made by Bach [2] that processes are analysis results of the section 3, which contains good prac-
useful, but not the central elements for the success of soft- tice raised in postmortems. In evaluating, for example, the
ware projects. The central point is the man - the hero who project Cel Damage, who took 85% of good practice, we ob-
solves ambiguous problems, which distinguishes between an served a lower incidence of problems, with only 20% of re-
expressed need and the one which will actually make the ported problems. Are there a linear correlation between the
client fully satisfied [2]. Moreover, it is clear the persistence number of problems encountered and good practices adopted
and belief in the success of teams who intensely focused on in game projects?
the software product.
For this analysis, was prepared to table 3, which is made up
Another interesting point is that management practices are of the game projects analyzed, the percentage of problems
a clear problem in the games industry. It does not reach found and the percentage of good practices adopted.
50% the number of projects in which a defined process
is adopted and only 25% (5 of 20) of the projects adopted The management practices are a clear deficiency in the game
good management practices. Moreover, only 40% of the industry. Not reach 50% the number of projects in which
projects used practices of quality control and was 10% (2 to we identified the adoption of a defined process of work
12
Belief in the
Creativity Focus on the Using Yesple Programming Good practices Continuous % good
Game Qualified team success of the Version control Agile modeling Defined process Quality control Feedback quickly Total
stimulus product tools good practices of management integration practices found
project
Beam Runner Hyper Cross Yes Yes Yes Yes Yes Yes Yes Yes Yes No No No No 9 69%
Gabriel Knights Yes No No No Yes No No No No Yes No No No 3 23%
Black & White Yes Yes Yes Yes No No No Yes No Yes No No No 6 46%
Rangers Lead the Way No No No No No Yes Yes No No No No No No 2 15%
Wild 9 Yes Yes Yes Yes Yes Yes No Yes No No No No No 7 54%
Trade Empires Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes No 12 92%
Rainbow Six Yes Yes Yes Yes Yes No No No No No No No No 5 38%
The X-Files Yes Yes Yes Yes Yes No No No No No No No No 5 38%
Draconus Yes Yes Yes No No No Yes No No No No No No 4 31%
13
Cel Damage Yes Yes No Yes Yes Yes Yes No Yes Yes Yes Yes Yes 11 85%
Command and Conquer: Tiberian Sun Yes Yes Yes Yes No Yes Yes Yes Yes No No No No 8 62%
Asheron's Call No Yes Yes Yes Yes Yes Yes No Yes Yes No No Yes 9 69%
Age of Empires II: The Age of Kings Yes Yes Yes No Yes Yes No No No Yes No Yes No 7 54%
Diablo II Yes Yes Yes Yes Yes No Yes No Yes Yes Yes No No 9 69%
Operation Flashpoint Yes Yes Yes Yes Yes Yes Yes Yes Yes No Yes No No 10 77%
Hidden Evil Yes Yes Yes Yes Yes No No No No Yes No No No 6 46%
Table 1: Occurrence of good practices in projects
Resident Evil 2 Yes Yes No Yes No Yes No Yes Yes No Yes Yes No 8 62%
Vampire: The Masquerade Yes Yes Yes Yes No Yes Yes Yes Yes No No No No 8 62%
Unreal Tournament Yes Yes Yes Yes No No Yes Yes No No Yes Yes No 8 62%
Tropico Yes Yes Yes Yes Yes Yes No No No No No No No 6 46%
Occurrences 18 18 16 16 13 12 11 9 9 8 6 5 2 143 7,2
% 90% 90% 80% 80% 65% 60% 55% 45% 45% 40% 30% 25% 10% 0,55 55%
Qualified team 90%
Continuous
integration 10%
0 2 4 6 8 10 12 14 16 18
16
14
Based on studies done by [34], [16] and [13] we believe that Gamasutra - The Art & Business of Making Games,
the adoption of agile practices in development of games can http://www.gamasutra.com/features/20020313/
achieve promising results. The agile practices seem to in- kreimeier\_01.htm, March 2002.
clude the best features of the game industry, as multidisci- [18] M. Marchesi, G. Succi, D. Wells, and L. Williams.
plinarity and the difficulties in modeling aspects like user ex- Extreme Programming Perspectives. Addison Wesley,
perience, the fun and enjoyment, whereas empirical methods 2002.
are more adaptive and react better to changes during project [19] M. McShaffry. Game Coding Complete. Paraglyph
implementation. We hope that this paper is a contribution Press, Scottsdale, 2003.
to increasing adoption of agile practices in the process of [20] M. Myllyaho, O. Salo, J. Kääriäinen, J. Hyysalo, and
game development. J. Koskela. A review of small and large post-mortem
analysis methods. In IEEE France, Paris, November
2004. 17th International Conference Software &
6. REFERENCES Systems Engineering and their Applications.
[1] S. W. Ambler. Modelagem Ágil. Bookman, São Paulo,
[21] F. Petrillo, M. Pimenta, F. Trindade, and C. Dietrich.
2004.
What went wrong? a survey of problems in game
[2] J. Bach. Enough about process: what we need are development. ACM Computer in Entertainment, CIE:
heroes. IEEE Software, 12(2):96–98, março 1995. 7(1), 2009.
[3] K. Beck. Extreme Programming Explained: Embrace [22] M. Poppendieck and T. Poppendieck. Principles of
Change. Addison-Wesley Professional, Reading, MA, lean thinking. 2003.
1st edition, 1999.
[23] R. S. Pressman. Engenharia de Software.
[4] E. Bethke. Game Development and Production. McGraw-Hill, São Paulo, 6th edition, 2006.
Wordware Publishing, Plano, 2003.
[24] R. Rouse. Game Design: theory & practice. Wordware
[5] A. Birk, T. Dingsoyr, and T. Stalhane. Postmortem: Publishing, Inc, 2001.
never leave a project without it. IEEE Software, 19,
[25] R. Rucker. Software Engineering and Computer
maio/junho 2002.
Games. Addison Wesley, December 2002.
[6] J. Blow. Game development: Harder than you think.
[26] E. Schaefer. Postmortem: Diablo ii. Gamasutra - The
ACM Press Queue, 1(10):28–37, February 2004.
Art & Business of Making Games, outubro 2000.
[7] D. Callele, E. Neufeld, and K. Schneider.
[27] B. Schofield. Embracing fun: Why extreme
Requirements engineering and the creative process in
programming is great for game development.
the video game industry. In 13th IEEE International
Gamasutra: The Art & Business of Making Games,
Conference on Requirements Engineering, August
March 2007.
2005.
[28] F. Shull, M. Mendonça, V. Basili, J. Carver, J. C.
[8] A. Cockburn. Agile Software Development. Addison
Maldonado, S. Fabbri, G. H. Travassos, and M. C.
Wesley Longman, Reading, MA, 1st edition, 2000.
Ferreira. Knowledge-sharing issues in experimental
[9] D. Cook. Evolutionary design - a practical process for software engineering. Empirical Software Engineering,
creating great game designs. GameDev.net, janeiro 9(1-2):111–137, 2004.
2001.
[29] F. Shull, J. Singer, and D. I. Sjøberg. Guide to
[10] K. Flood. Game unified process. GameDev.net, Maio Advanced Empirical Software Engineering.
2003. Springer-Verlag, London, 2008.
[11] J. P. Flynt and O. Salem. Software Engineering for [30] I. Sommerville. Software Engineering. International
Game Developers. Software Engineering Series. Course computer science series. Addison-Wesley, London, 6th
Technology PTR, 1st edition, November 2004. edition, 2001.
[12] A. Gershenfeld, M. Loparco, and C. Barajas. Game [31] T. Stalhane, T. Dingsoyr, G. K. Hanssen, and N. B.
Plan: the insider’s guide to breaking in and succeeding Moe. Post Mortem? An Assessment of Two
in the computer and vieo game business. St. Martin’ s Approaches, chapter Empirical Methods and Studies
Griffin Press, New York, 2003. in Software Engineering, pages 129–141. Springer
[13] A. Gibson. Agile game development and fun. Berlin / Heidelberg, 2003.
Technical report, University of Colorado Department [32] F. Tsui and O. Karam. Essentials of software
of Computer Science, 2007. engineering. Jones and Barlett Publishers, São Paulo,
[14] W. Hamann. Goodbye postmortems, hello critical 6th ed edition, 2007.
stage analysis. Gamasutra - The Art & Business of [33] C. Wohlin, M. Höst, and K. Henningsson. Empirical
Making Games, julho 2003. Methods and Studies in Software Engineering, chapter
[15] J. Highsmith and A. Cockburn. Agile software Empirical Research Methods in Software Engineering,
development: The business of innovation. IEEE pages 7–23. Springer Berlin / Heidelberg, 2003.
Computer, 34:120–122, 2001. [34] S. Xu and V. Rajlich. Empirical validation of
[16] L. C. C. Kasperavicius, L. N. M. Bezerra, L. Silva, test-driven pair programming in game development.
and I. F. Silveira. Ensino de desenvolvimento de jogos Computer and Information Science, 5th IEEE/ACIS
digitais baseado em metodologias Ágeis: o projeto International Conference on, 0:500–505, 2006.
primeira habilitação. In Anais do XXVIII Congresso
da SBC - Workshop sobre Educação em Computação,
pages 89 – 98, Belém do Pará, Julho 2008.
[17] B. Kreimeier. The case for game design patterns.
15
View publication stats